Pixelgrammar Lab

Room to question.
Space to experiment.

Working experiments in AI, simulation and creative technology. Explore the systems we build, the questions behind them and what we learn along the way.

Explore the use cases

Abstract futuristic form exploring the meeting point of AI and design

Technology in practice

Ideas you can
interact with.

Built by Pixelgrammar.
Two working experiments to explore.

01 / Simulation & typed AI

LUS–222

Experimental simulation

What changes when a human and an AI share the controls?

What we built: A browser-based 3D flight simulator over Porto, connecting TypeSafe’s JEV to aircraft telemetry and structured decisions. Human and AI control share the same instruments, with manual takeover available.
Where it can lead: Interactive simulation, decision-support interfaces and systems where people need to inspect, supervise and override AI decisions.

TypeSafe / JEV
3D & WebGL
Human oversight

Explore the simulator

Research prototype with illustrative physics, not certified flight training.

02 / Autonomous agents & data

BotFfett

Simulated capital

How do we make autonomous decisions observable?

What we built: An autonomous trading experiment that connects market analysis, risk sizing and simulated execution. A public dashboard exposes the portfolio, working theses and decision log, including decisions to take no action.
Where it can lead: Agent workflows, connected data products and monitoring tools that help teams understand what a system did, and why.

AI agents
Data integration
Decision monitoring

Explore the experiment

An experiment using simulated capital. The capture is a dated snapshot, not a live result or a performance promise.

Where curiosity
becomes a prototype.

Four connected fields.
Ideas tested in practice.

Content & agentic workflows

AI agents, orchestration and A/B learning that connect creative work with repeatable processes.

Robotics & Physical AI

Sourcing, integration and pilots that put intelligent systems to work on real-world tasks.

Smart devices & edge AI

Embedded intelligence, interoperable devices and privacy considered from the first prototype.

Data & insight products

LLM applications, analytics loops, KPI dashboards and decision support for teams.

How we work

A clear question.
A measurable next step.

Every experiment starts with a hypothesis, defined KPIs and a decision about when to stop or go further.

01
Discover
Define the question, hypothesis and KPIs.
02
Build
Turn the hypothesis into a focused prototype.
03
Pilot
Test in a live setting with clear risk limits.
04
Scale
Roll out what the evidence supports.
05
Learn
Use feedback to improve, stop or double down.

Our principles

Open-source first
Privacy by design
Vendor neutral
Measurable impact
Fast iteration
Ethical use

An experiment in digital creativity

New.Vogue.LabFashion, reimagined.

A space to explore fashion, luxury and visual culture through AI. We experiment with imagery, content workflows and new creative possibilities — then share the results.

Follow New.Vogue.Lab

White dog in an imaginative fashion editorial

Fashion portrait blending a model with sculptural forms

AI fashion exploration with a luxury-inspired ice cream motif

Inside New.Vogue.Lab

Beyond the image.
Inside the experiment.

Fashion and luxury are the creative setting. Content, automation and growth are the questions we are testing.

The objective

Build a coherent luxury aesthetic and consistent storytelling, while testing workflows that encourage meaningful engagement, saves, shares and qualified enquiries.

The hypothesis

Combine visual guardrails for palette, lighting and textures with AI look ideation, scheduling and a small CRM. Can we create more cohesive content, faster, and learn from better signals?

01Creative engineFrom brand DNA to a consistent visual world.

Brand DNA prompts: Colour, fabric, silhouette and backdrops define the visual direction.
Look ideation: Seasonal capsules and editorial themes guide each new exploration.
Visuals & motion: Short-form video and imagery share a consistent approach to lighting.
Caption craft: Tone-controlled copy carries the story to a fashion and luxury audience.

02Automation stackThe workflows behind the creative output.

Content calendar: Themes, publishing windows and A/B variants keep the experiment organised.
Reels workflow: Cuts, beat matching and subtitles come together in a connected production process.
Social listening: Comment classification, direct-message routing and saved replies help organise conversations.
Micro-CRM: Tag interested prospects, capture intent signals and record look requests.

03What we measureSignals that help us decide what to try next.

Brand: Follower quality, saves, shares and mentions.
Content: Retention in the first three to five seconds, watch percentage and carousel depth.
Growth: Profile visits, link clicks, direct-message conversations and subscription interest.
Commerce signals: Appointments, sample loans and look requests, without treating interest as a completed sale.

Ideas,
in motion.

Four short studies in AI and visual storytelling.

Motion TV

A wall of televisions becomes a study of eyes, screens and analogue texture.

Silent study · 5 seconds

Black Pantera

A fashion portrait exploring presence, movement and a dark editorial mood.

Silent study · 5 seconds

Photo Shooting

A moving collage of close-ups and fashion photography.

Silent study · 5 seconds

Fashion Prototype

An experimental fashion portrait, brought from a still image into motion.

Silent study · 5 seconds

What’s next

Room to question.
Space to experiment.